reinventing dsl with copper pair bonding - · pdf filereinventing dsl with copper pair bonding...
TRANSCRIPT
Technology Forum at TELECOM ‘05
Reinventing DSL Reinventing DSL with Copper Pair Bondingwith Copper Pair Bonding
Wednesday, October 261:30 – 3:00 p.m.
Venetian Ballroom I
Reinventing DSL with Copper Bonding
ATIS Technology ConferenceTelecom 05
October 26, 2005
Reinventing DSL: Higher Speeds
Competition– Cable– 3G Wireless
Migration– From 1.5 Mbps to 15 Mbps or more
Standards– Need cost-effective, standard solutions
Bonding Defined
Inverse Multiplexing– Payload Bonding– Combining the payloads of 2 or more DSL links into a
single “fat pipe”– Many Standards available
• MLPPP (RFC 1990) – relies on underlying HDLC• IMA – Inverse Multiplexing for ATM• T1.427, G.998
– .01/.1: ATM Cell Bonding– .02/.2: Ethernet Transport (64/65 TC) – .03/.3: Time Division Inverse Multiplexing (TDIM)
• IEEE 802.3ah – Ethernet bonding with SHDSL/VDSL
What Bonding with What DSL?
IMA
64/65 Ethernet Bonding
MLPPP
ATM Bonding
TDIM
ADSL/ ADSL2+
VDSL1/2
SHDSL
DS-1
Theoretically, any bonding
works with any DSL
What Bonding with What DSL?
IMA
64/65 Ethernet Bonding
MLPPP
ATM Bonding
TDIM
ADSL/ ADSL2+
VDSL1/2
SHDSL
DS-1
In practice, these mappings are the most common
Beyond Standards
Multiple Input Multiple Output (MIMO)– Can be applied mostly independent of payload
bonding method– Rates can be boosted through
• In-Domain Crosstalk Cancellation• Out-of-Domain Crosstalk Compensation• Pair-diverse Coding
– Out of domain gains can vary widely• Most gain requires large (8-12 pairs) group size• Depends on # and type of other system and noise
Panelists
Jürgen Lison, Alcatel– Bonding and ADSL2+
Rouben Toumani, Ikanos– Bonding and VDSL2
Gary Tennyson, BellSouth– Bonding Deployment Considerations
All rights reserved © 2005, Alcatel
Reinventing DSL with Bonding
Jürgen Lison, Alcatel
Reinventing DSL – USTA 2005 | Page 2 All rights reserved © 2005, Alcatel
Topics
Why bond?
Capture new customers with extended reach
Address business needs with symmetrical services
Offer triple play services
Shannon’s Law and market realities
Bonded ADSL2+ fills the gaps
Bonding challenges and solutions
Summary
Reinventing DSL – USTA 2005 | Page 3 All rights reserved © 2005, Alcatel
How some people bond lines
Key features:Poor scalabilityHigh fixed investment costsSafety concernsAnd not very flexible
Reinventing DSL – USTA 2005 | Page 4 All rights reserved © 2005, Alcatel
Why Bond ADSL2+?
At any REACH – Double RATE (DS, US)
Generally Available and shipping TODAY
Stable, standard technology
Chip synergies with legacy ADSL
Greater range of marketable bandwidths than single pair VDSL/VDSL2
Reinventing DSL – USTA 2005 | Page 5 All rights reserved © 2005, Alcatel
Expanding the customer (and revenue) base for DSL
Extending ReachBonded ADSL2+ doubles data rates at any given reachCustomers previously unable to qualify for DSL – now qualify
Symmetrical ServicesDouble upstream bandwidthT1 replacement opportunityTarget SMB customersCompetitive positioning over cable
Reinventing DSL – USTA 2005 | Page 6 All rights reserved © 2005, Alcatel
Delivering the Triple Play
Reinventing DSL – USTA 2005 | Page 7 All rights reserved © 2005, Alcatel
How much bandwidth is needed at the home?
HDHD
Depends on the services offeredVideo
High bandwidth and high QOSHD and SDMultiple STBs, DVRs, and PIPs
VoiceLow bandwidth but high QOS
DataHigh bandwidth but burstySome QoSSupply-side product: “If you build it, they will come.”
SDSDSDSD
TV 1
HSIHSITier 1
TV 2DVR 1
HDHD DVR 2
Data20
Mbps
HSIHSITier 2
“Table Stakes”
StillMoreData
HSIHSI
MoreData
SDSD TV 3
100+Mbps
40Mbps
Triple P
lay Marke
t
Reinventing DSL – USTA 2005 | Page 8 All rights reserved © 2005, Alcatel
IPTV and interactivity
IPTV drives interactive servicesMultiple Picture-in-picture streamsOnline gamingVideo on Demand
Services drive more bandwidthMultiple HD streamsQuick response video games
Competition drives more bandwidthCable is and will continue to differentiate on “perceived” data bandwidth
Order Print^
The market requires an access network transformation
Reinventing DSL – USTA 2005 | Page 9 All rights reserved © 2005, Alcatel
0
10
20
30
40
50
60
70
80
0 1,000 2,000 3,000 4,000 5,000 6,000
Loop Length (ft)
Dat
a R
ate
(Mbp
s)
ADSL2+
ADSL
Better data rates are possible
Basic video can be delivered, but HDTV and any cable competition hurts marketability
DLC loop lengths are not IPTV-enabled with ADSL2+
Single Pair ADSL or ADSL2+ won’t deliver the Triple Play
TriplePlay
Bandwidth
Triple Play Market
FTTC FTTN
Reinventing DSL – USTA 2005 | Page 10 All rights reserved © 2005, Alcatel
0
10
20
30
40
50
60
70
80
0 1,000 2,000 3,000 4,000 5,000 6,000
Loop Length (ft)
Dat
a R
ate
(Mbp
s)
VDSL2
ADSL2+
ADSL
Enables complete triple play over copper
Simple approach:One pair Easy management
New technologyEconomics of scaleHigher impact from interference
VDSL2 is one solution
TriplePlay
Bandwidth
Triple Play Market
FTTC FTTN
Reinventing DSL – USTA 2005 | Page 11 All rights reserved © 2005, Alcatel
Bonded ADSL2+ transforms the copper access plant
Provides the widest range of options for Triple PlaySynergies with existing ADSL
ChipsetsPower profile
ComplexitiesMultiple pairs not available everywhereManaging data across two pairsB/OSS issues to be resolved
0
10
20
30
40
50
60
70
80
0 1,000 2,000 3,000 4,000 5,000 6,000
Loop Length (ft)
Dat
a R
ate
(Mbp
s)
VDSL2
Bonded ADSL2+
ADSL2+
ADSL
FTTC FTTN
TriplePlay
Bandwidth
Triple Play Market
Reinventing DSL – USTA 2005 | Page 12 All rights reserved © 2005, Alcatel
Execution presents challenges…that have solutions!
Challenges
Finding additional copper pair between subscriber and DSLAM
New “Bonding-ready” CPE required
OSS systems not currently supporting bonding
Solutions
Most homes have two pair for legacy POTS requirements
System level bonding eliminates DSLAM rewiring: use the next pair, regardless of position
IPTV: New CPE will be required regardlessERDSL: New customers need CPE
OSS is already designed to support two lines per customers
Element management tools enable efficient bonding management within current operations practices and systems
Reinventing DSL – USTA 2005 | Page 13 All rights reserved © 2005, Alcatel
Integrated bonding with service provider’s NMS/OSS systems
Operations procedures are the SAME as currently employed
TL1 upstream interface for flow through provisioning, performance monitoring, and alarm surveillanceTools to examine equipment inventory and customer statusEasy to use GUI for troubleshooting, and maintenance purposes
Simplicity of system-level bondingOSS does not have to worry about “which two lines to pair”“Any line” to “any line” capabilityIntegration into existing element management
Reinventing DSL – USTA 2005 | Page 14 All rights reserved © 2005, Alcatel
Summary
Bonding Copper PairsEnables triple play services over copper and node/curb distancesProvides means to expand residential DSL customer baseAllows competitive SMB service offerings
Operational challenges have clear solutions
All of this is available – TODAY
Reinventing DSL – USTA 2005 | Page 15 All rights reserved © 2005, Alcatel
www.alcatel.com
Ikanos Confidential
Reinventing DSL:Delivering 25Mbps & Beyond Over Existing Copper PairsRouben Toumani, PhD. Ikanos Communications
2Ikanos Confidential
Overview
» The copper access network is evolving driven by new applications and services
» DSL technologies have evolved to meet the need– ADSL -> ADSL2+ -> VDSL -> VDSL2– VDSL2 has emerged as a universal DSL technology to serve multiple
deployment options in the access network– From Dial-up to Fiber-Fast Broadband Performance– From FTTP to Exchange-Based distances
» Promising new technologies on the horizon– Bonding of two or more pairs of copper
• Great performance enhancement with existing technologies• Applies to multiple physical layers (ADSL to VDSL2)
– Dynamic Spectrum Management (DSM)– Noise cancellation techniques (MIMO)– Dynamic Noise Management
3Ikanos Confidential
MxU, FTTNMxU, FTTN
50-100 Mbps50-100 Mbps
Fiber/Copper Deployment Scenarios
Node
Remote
CO, RTCO, RT
20-50 Mbps20-50 Mbps
CentralOffice
<1Mile<1Mile
11--2 Miles2 Miles
>2 Miles>2 Miles
FTTB FTTB –– 1Kft1Kft
FTTHFTTH
4Ikanos Confidential
Telco Access Evolution
1995 2000 2005
56K/28K NarrowbandDialup
NarrowbandDialup
• Basic web access• Email
8M/800K Gen 1 (ADSL)Broadband
Gen 1 (ADSL)Broadband
• Enhanced web access• Streaming audio
100M/100M,25M/5M
VDSL, VDSL2Broadband
VDSL, VDSL2Broadband
• Interactive Broadband• Triple-Play +
Gen 2 (ADSL2+)Broadband
Gen 2 (ADSL2+)Broadband
• Streaming audio and video• Enhanced peer-peer sharing
25M/2M
4 kHz
1.1 MHz
2.2 MHz
30 MHz
5Ikanos Confidential
Interactive Broadband: The Next Wave
12 Mbps/channel HDTVInteractive Features,
Personalized Services
Media-rich,Multi-player
Gaming
WidebandVoIP 4x-10x Faster
Web Access,Remote Access
of Home PC
Video Blogging, Video Upload,
Digital Photo Upload
4 Mbpsbi-directional
Enhanced Video + Audio, Video Surveillance
6Ikanos Confidential
VDSL2: The Universal DSL Technology» Cost-effective fiber bandwidth over copper
– Capability to deliver 100 Mbps symmetric services
– Supports triple-play
– Supports interactive services (enhanced upstream)
– Designed for integration with IP packet based networks
» Built on solid, proven DMT technology
– 100 Million lines and counting!
– Seamless multi-mode (ADSL/ADSL2/ADSL2+/VDSL/VDSL2)
» Key technology enhancements
– Wider spectrum utilization up to 30 MHz
– Trellis Coded Modulation (TCM)
– Deployment-optimized implementations: Profiles
– PSD shaping for co-existence with the deployed services
– Optional U0 band expanded to 276 kHz for long reach
7Ikanos Confidential
Technology EnhancementsBonded VDSL2
» VDSL2 Copper Pair Bonding– Standards (ITU-T G.998 / ATIS T1.427 series) compliant
• Packet (Ethernet): Seamless integration into IP based networks
• Also ATM for legacy networks and TDIM for DS3 transport
– Data aggregation protocol above the Physical Layer• Offers greatest flexibility / optimized implementations
» Market Drivers for Bonded VDSL2– Ensure ubiquitous service offering / geographic footprint
– Commercial MTU’s, Residential MDU’s• Business: High reliability, guaranteed symmetric rates
• Avoids copper ownership / building access issues
– Single Family Residential• Consumer: Offer the highest broadband rates available
• Perfect for deep fiber / PON network architectures
8Ikanos Confidential
Bonding 2 pairs of VDSL2
VDSL2 Profile 12a
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0
0 1,000 2,000 3,000 4,000 5,000 6,000
Loop Length (feet)
Mbps
Downstream UpstreamDownstream (2 pairs bonded) Upstream (2 pairs bonded)
•12 self-disturbers (99% worst case)
•24 AWG
9Ikanos Confidential
Technology Enhancements: DSM (Dynamic Spectrum Management)
» Four Levels of DSM
– Level 0: Nothing “hogs” the available wire line spectrum
• In use today: Spectrum Management Rules (T1.417)
• Systems shall not cause each other “significant degradation”
– Level 1: Politeness = Near transmitter “speaks softly”
• Transmits only the power needed for service offering
• Especially useful for “near-far” problem
– Level 2: Spectrum Balancing
• Adaptively determines PSDMASK
• Shorter loop uses higher frequencies with less penalty
– Level 3: Vectoring = Intelligent Service Provider Network
• Active management of both CO and CPE modems
• Serious number crunching gains serious performance
10Ikanos Confidential
Technology Enhancements: MIMO (Multiple Input Multiple Output)
» Technology used in wireless networks
» Now being introduced for copper bonded systems– Intra-network transport (Cell site backhaul, T1/T3 gap)
– 12+ pairs bonded in end-to-end system, vector managed
» The ultimate Level 3 DSM– Vectoring = true mitigation of NEXT and FEXT crosstalk
– MIMO NEXT mitigation: US is easier• Co-located receivers seeking common enemy – exploits the
alien crosstalk correlation across lines
• Useful to create greater rate symmetry (better US)
– DS more complex: • Must use transmitter pre-compensation
• Pre-comp is based on data of varying accuracy
11Ikanos Confidential
Summary
» Drivers of copper access network evolution include:– Triple Play / HDTV delivery
– Peer-to-peer applications requiring greater upstream
» DSL technologies have evolved to meet the need– Bonding provides additional coverage
– DSM & MIMO: active noise mitigation is the next step
» VDSL2: – VDSL2 has arrived as the universal DSL technology
– Ready to implement future technology enhancements
1
Reinventing DSL with Copper Pair Bonding
October 26, 2005
Gary TennysonBellSouth Telecommunications
2
Building a Competitive Broadband Network
• New distribution areas – Fiber To The Curb (FTTC)
• Base of existing distribution areas –Fiber To The Node (FTTN)
• Both drive fiber deeper into the loop network
3
Fiber To The Curb
• Telephony and DSL service is supported, not an overlay
• Fiber placed throughout a serving area as it is developed
• Optical Network Units (ONU) provide for conversion from an optical to an electrical interface– Maximum Distance from ONU to a subscriber’s
residence/business is 500 feet– Short drop allows us to take advantage of the
very high data rates offered by VDSL
4
Fiber To The Node
• New platforms will support the existing base of DSL lines– Large existing base of remote DSLAM’s could
create a complex spectrum management problem in an overlay network
– Multi-mode ports to support existing DSL• Initial focus on ADSL2plus• Planning to use VDSL2 as we move forward• Bonding
5
What is Bonding ?
• Provides for summing the capacity of two or more DSL lines– Involves multiple physical-layer transceivers– One data stream ‘seen’ by the user
• Provides for greater data rates and/or operation on longer loops
• Needed with ADSL2plus for data rates greater than about 12 – 18 Mbps in a multi-pair environment
• Not unique to ADSL2plus
6
Why Choose Bonded DSL ?
• Decision requires insight into several areas:– DSL Rate/Reach– Distribution of loop lengths– Availability of vacant pairs
• What is the targeted maximum data rate ?• What is the targeted coverage, i.e., what
percent of loops should be capable of supporting the targeted data rate ?
7
DSL Reach Depends on the Targeted Data Rate
• Need to consider different disturber types– Repeatered T1, for example– Self-FEXT not typically seen as a limiting factor
at lower data rate and frequencies; more impact at higher data rates and frequencies
– Existing DSL transferred to new platform –common launch level
• Bridge Taps• Gauge – coarse gauge may not provide the
improvement that you would expect
8
Distribution of Loop Lengths
9
Loop Reach Impacts Costs
• Shorter loop reach => More nodes– Per site costs
• Cabinets• Power• Interconnection• Easement costs
– Are interconnection points available ?• Existing DLC RT’s – power, pad, and interconnection
are available• Existing FDI’s – interconnection is available• Re-sectionalized plant
• Bonding provides tradeoff between pair availability and the number of nodes
10
Planning Considerations
• Unequal data rates– Two DSL’s cannot simultaneously be
experiencing the worst-case crosstalk– Bonding will support unequal data rates to take
advantage of this• Are vacant pairs available in the serving
terminals ?• Two lines are needed from the DSLAM to
the CPE modem– New drop will often be needed (buried/aerial)– New Inside Wire may be needed – Truck roll
11
Operational Considerations
• System-wide bonding vs. card-level bonding– System-wide bonding allows bonding of a new
line to an existing line without a transfer– Card-level bonding can require a transfer
• The second line may not have an underlying POTS line– Loop testing without a telephone number ?– Sealing Current ?
• Bonding-capable CPE is required